Kohir has long been the head of meta research as Joel Panio as his chief AI officer.

Investors once watched Canada’s AI startup as a promising contender to challenge Open and Entropic in the race to make Kohir a Frontier AI model, its backers put about $ 1 billion on their condition on CEO Eden Gomez, which was shared on Google as Google.

But Kohir’s AI models have fallen behind the latest art, and its business has not measured it like its rivals.

Now, the company is bringing an experienced research leader to improve its AI’s efforts: Kohir hired former AI research meta VP, Joel Panio, who had previously monitored the basic AI Research (Fair) lab of Tech Dev. In the role of his newly created Chief AI officer, Panio will oversee the AI strategy in Kohir’s research, products and policy teams.

Panio, a Canadian AI scientist and a professor of McGill, helped Yan, a pioneer of neural networks, to guide the initial development of Open Lama AI models of Meta. Panio left Meta in May about eight years after the company.

This is a huge fare for Kohri, and he is helping the veteran further investigating his hopes, improves its research and product pipeline, and recruits high capabilities.

Fare comes at a significant moment for Kohir: Million is allegedly trying to collect up to 500 million in Company A .3 6.3 billion value – An inspirational amount was not competing with the choice of Openi, Google, Meta and Anthropic, whose combat breasts cost billions of billions.

But when its rivals are trying to develop (or exceed or exceed human performance on a variety of tasks, Koheer’s focus is a narrow attention. Startup mainly produces AI applications that can solve practical issues for businesses and government agencies, emphasizing privacy and security.

In an interview to Tech Crunch, Panio said that focusing on Kohir’s real world, enterprise applications are something that he is excited about. “Many players there focuses on AGI, sprintles and so on,” said Panio, referring to companies like his ex -employer, Meta. “They do not necessarily know what is being used for this AI.”

It pointed to the launch of Openi’s GPT5 last week, which many people felt like they were disturbing, as proof that the AGI acquisition timeline “could be a little longer than we think.” In the meantime, Panio says there is a lot of scope for more practical AI models to jump in productivity in different industries.

Panio, a Canadian resident, said his eyes were on Kohri since his foundation was laid in 2019, and he is excited to contribute to the company whose founders are based in his native country.

Beyond patriotism, Panu feels that opportunity with Kohir is a good opportunity to travel beyond research. At the festival, Panio worked on the projects monitoring the projects that could take anywhere from 18 months to 10 years. Now, they will work with more tough timelines as well as joining consumers and products. And although Kohir has less resources than meta, Panio said she would be more firm in her new role.

Kohir’s latest product is an AI agent platform called North, which business and government agencies can privately deployed on their infrastructure, which is an attractive idea for many of its users, who are bank and federal organizations that handle highly sensitive data. This is accompanied by open source providers competing against depressic and meta, whose models can also be run locally, but at a lower price. Kohir is conditioning that it can defeat the open model, offering more help around his private deployment.

Panio said he was particularly interested in the preparation of more and more research in the north, to detect ways to manufacture AI agents in private and secure settings and to make standards to evaluate these systems. Panio also said that he was interested in finding out how AI agents in the real world interact with each other.

A quick challenge for Panio will replace Koheer’s AI research, Sarah, who announced it. Departure After several years of helping the company’s research program this week. Given the demand for AI’s capabilities, it can be difficult to hire the AI researcher in the current market by being in the current market.

But Panio has seen it as an opportunity to “bring a lot of skills”, saying that when he left the Meta, several former colleagues advised that he would follow him in a new AI lab. However, he emphasized that Kohir has a solid base for AI researchers, and it is important not to bring only anyone.

“It is not necessary to hire a group of superstars to create a superstar team,” said Panieu. “Really about how people work together.”

Of course, Meta’s AI unit looks very different today when Panio was there just a few months ago. During the summer, Mark Zuckerberg received a recruitment facility, the industry’s best AI researchers to join the MSL were allegedly presented to the North of 100 million Million. This also forced its star employees to increase the compensation, thus making it difficult for the smaller players to get to top AI researchers.

As Meta, Openi and Anthropic throw billions of dollars at their AI’s efforts, Kohir is trying to minimize. That is, it means to bet the calculation research – a type of type that can quickly turn into a compulsive product and keep the company in the race.